A conventional decision feedback equalizer (DFE) uses prior symbol decisions (i.e., hard decisions) to remove inter-symbol interference (ISI) from future symbols. A DFE includes a linear feed forward filter to filter samples and a linear feedback filter to weight hard decisions of prior symbols. The weighted hard decisions are subtracted from the output of the feed forward filter to provide equalized samples or soft decisions. Tap weights of the feed forward and feedback filters may be controlled to reduce error between the soft and hard decisions.
Erroneous hard decisions propagate through the DFE feedback filter to corrupt future symbols. A DFE may include an erasure slicer to mitigate such error propagation by avoiding feedback of unreliable (i.e., potentially erroneous) hard decisions, referred to as “erasing.” A hard decision may be identified for erasure when a corresponding soft decision is within a threshold distance of a decision boundary. The threshold may be referred to as an erasure threshold or an erasure slicer threshold.
The erasure threshold may impact DFE performance. If the threshold is too high, reliable hard decisions may be unnecessarily erased. If the threshold is too low, unreliable hard decisions may propagate errors to future symbols.
Multi-dimensional erasure threshold optimization techniques have been proposed, but are not practical for cost-sensitive, real-time applications.
In the drawings, the leftmost digit(s) of a reference number identifies the drawing in which the reference number first appears.